Hyperspectral system trade-offs for illumination, hardware and analysis methods: a case study of seed mix ingredient discrimination | Zendy

Carolina Blanch-Pérez del Notario | Zendy; Carlos López-Molina | Zendy; Andy Lambrechts | Zendy; Wouter Saeys | Zendy

Open Access

Hyperspectral system trade-offs for illumination, hardware and analysis methods: a case study of seed mix ingredient discrimination

Author(s) -

Carolina Blanch-Pérez del Notario,

Carlos López-Molina,

Andy Lambrechts,

Wouter Saeys

Publication year - 2020

Publication title -

journal of spectral imaging

Language(s) - English

Resource type - Journals

SCImago Journal Rank - 0.256

H-Index - 6

ISSN - 2040-4565

DOI - 10.1255/jsi.2020.a16

Subject(s) - hyperspectral imaging , computer science , artificial intelligence , computer vision , image processing , pixel , segmentation , convolutional neural network , image resolution , pattern recognition (psychology) , image (mathematics)

Block size CNN feature size Mean accuracy Minimum accuracy Iterations required Convergence time (s) 1 × 1 9 73.3 % 50.1 % 35 90.0 3 × 3 9 × 9 = 81 84.6 % 68.0 % 35 87.0 5 × 5 25 × 9 = 225 87.9 % 75.7 % 35 94.3 7 × 7 49 × 9 = 441 89.1 % 74.3 % 25 75.6 9 × 9 81 × 9 = 729 90.0 % 76.2 % 25 101.0 11 × 11 121 × 9 = 1089 89.4 % 71.0 % 25 119.2 Table S1. Impact of input image block size on CNN performance (pixel accuracy), Snapscan image under halogen illumination.

The content you want is available to Zendy users.

Already have an account? Click here to sign in.

Having issues? You can contact us here

Accelerating Research